A Real Time Control System based on a Fuzzy Compiled Knowledge Base

This paper presents the predictability problem in a real-time expert control system. It is a special possibilistic expert system (FESPC), developed in order to focus on fuzzy knowledge. We are not dealing high with level reasoning methods, because we think that real-time problems can only be solved by rather low-level reasoning. When humans engage in problem solving, the qualitative aspects of knowledge are hierarchically organized to provide concept association and reasoning, based on fuzzy logic inference. Most of the overall run-time of fuzzy expert systems is used in the match phase. To achieve a fast reasoning the number of fuzzy set operations must be reduced. For this, we use a fuzzy compiled structure of knowledge, like Rete, because it is required for real-time responses. Solving the match-time predictability problem would allow us to built much more powerful reasoning techniques.